These days, one can hardly turn on a news programme without hearing about dire economic forecasts. Especially over the past year, as the impact of various stimulus packages and the pandemic fallout became more visible, economists everywhere were - and still are frantically ringing alarm bells.
Are their warnings a mere continuation or a resumption of the global economic downturn of a decade ago? Or is there something new to be alarmed about? That is where econometrics comes in.
Broadly put, econometrics is the math behind economic modelling. It looks at discrete variables within the aggregated data to winnow out what matters at the time of the data's analysis. In other words, econometrics provides documented, observable results of specific, economically related phenomena.
A perfect, recent example of such a phenomenon is the American housing bubble that resulted in the 2008 global financial crisis.
To understand econometrics' role in forecasting economic outlooks and steering economic policy, you have to know what econometrics is - and what it isn't.
That's what your Superprof looks at today.
Econometrics: A Brief Background
The concept and study of economics as we know it today has been around for centuries. Anyone who's ever cracked an Economics textbook knows that Adam Smith is widely considered the father of modern economics but the topic has been debated since antiquity.
By contrast, econometrics has only been around for about 100 years.
Pawel Ciompa, a Polish economist, coined the term 'econometrics' in 1910 but his contributions to the field were soon overshadowed by the two economists who are now referred to as the founding fathers of this discipline.
Dutch economist Jan Tinbergen got an early start working with statistics while rounding out his mandated civil service at the Central Bureau of Statistics (CBS) in The Hague. He went on to lead its newly-formed mathematical statistics and business surveys department, where he could indulge his passion for physics and mathematics, economics and politics.
With vast troves of data at his disposal, he began testing various theoretical models and, by the time the Dutch government created their Bureau for Economic Policy Analysis - of which he was the first director, his statistical models were influencing economic policy in Pakistan, Indonesia and the United Arab Emirates.
As Mr Tinbergen was developing his econometric models, Norwegian economist Ragnar Frisch was making the case for economics to follow the same empirical/theoretical path that other sciences, particularly physics. He reinforced his position with his foundational document, published in 1926, that emphasised cardinal and ordinal utility.
Messrs Frisch and Tinbergen arrived at the same conclusion by travelling separate paths. They both established that the collection, examination and analysis of economic data must be handled scientifically to reach provable conclusions rather than being made to fit abstract, theoretical concepts.
That idea revolutionised the practice of economics. From then on, conditions that drive economic phenomena were assessed in contrast to each other, proving over and over again that correlation is not causation. By modelling datasets, econometrists were finally able to prove that Condition X does not necessarily result in Outcome Y.
It was Mr Frisch who applied the term 'model' to economics; he also coined the terms macro- and microeconomics and made formal the Production Theory of economics.
In 1969, the men were jointly awarded the Nobel Memorial Prize in Economic Sciences. Theirs was the very first such award - until it was given to them, it didn't exist, and rightfully so. Until their work, economics was not considered a science.
Also, discover how, since that time, theoretical econometrics has evolved.
What Econometrics Does
If you think of economics/econometrics in terms of George Orwell's The Time Machine, economics would be the Eloi to econometrics' hard-working Morlocks. It's not likely that an econometrist will suddenly rise up to devour an economist, though.
For centuries, the field of economics was heavy on theory and exceedingly light on application. We get a clue of that in Adam Smith's 'Invisible Hand' postulate. Where does that hand come from and who does it belong to? Or should we simply put all economic phenomena down to fate and luck of the draw?
We find further proof of the previously fanciful nature of economic thought in the diverging opinions of Karl Marx and Thomas Malthus.
The latter's doomsday predictions of populations outgrowing the food supply led the former to deem that controlling the production (of food and other resources) was of paramount importance. Specifically, his focus was on who should control them.
The two men were in agreement in only one aspect: they both failed to consider variables that might change their predicted outcomes. Or, more generously, they were unaware of the variables that would influence their theories.
That's where econometrics comes in.
If Mr Malthus had been more aware of the technological advances that would greatly boost food production - things like agricultural machines and commercial bakeries, his forecast might not have been so glum. And if Mr Marx could only have reviewed empirical data that proved that capitalism is far more yielding than the rigid lines he painted it with, he might not have advocated for class wars.
By melding economic theory with real-time observation of economic activity and the conditions driving it, one can derive a quantitative analysis of the economic phenomenon in question.
How much more value could Either Marx or Malthus have added to the conversation if they had tried to model their conclusions?
Now delve deeper into the three goals of econometrics...
How Econometrics Works
Every single economist, even those predating Adam Smith has given us philosophical and sometimes elegant theories on how and why money works. In hindsight, though, they were more expressing opinions than offering concrete facts about market drivers and people's lives.
Econometrics drives economics; it's what gives all of those theories weight. After all, if economic theory is not quantifiable, its usefulness is fairly limited. So now do econometrists quantify economic data?
Linear multiple regression is most often used to estimate how much a change in one factor or variable impacts the economic aspect under scrutiny. Those two variables are labelled 'explanatory' and 'dependent', respectively.
So, if economists want to measure how raising taxes will affect consumer spending, for example, they would focus only on the relationship between those two variables, assuming that others that impact consumer spending, such as gross income and personal wealth, will remain constant.
Keeping that in mind, here are the four steps that econometrists take to quantify their answer:
- establish a hypothesis
- outline the parameters; select the variables to be examined
- specify the statistical model that best reflects the proposed econometric theory
- linear models are by far the most common; a change in one variable should represent a similar change in the other variable
- add a 'catchall' to the dependent variable to account for any unknown determinants
- plug all of the variables into your econometric software platform
- verify that the results make economic sense by testing the hypothesis
Establishing hypotheses is generally the purview of theoretical econometrists but proving and testing them is what applied econometrics is for.
In our scenario above - how raising taxes will affect consumer spending, you might expect that a 1% tax increase might lower spending in economically disadvantaged populations but, the more wealth people have, the less their spending out be affected by such a small tax increase.
If the linear regression model bears that hypothesis out - and shows to what degree lower-income consumer spending would be reduced, governments might decide to revise their decision to tax the population across the board or even that the tax revenue won't offset the loss of consumer spending.
Criticisms of Econometrics
If you torture the data long enough, it will confess. - Ronald Coase
Probably everyone has - or, at least has heard an economic truism such as the one Mr Coase is widely reported to have uttered. My boss used to say 'The numbers are the numbers'; a rather obvious and nonsensical way to trivialise data revelations.
Still, there are plenty who add their voices to the body of criticism of econometrics; most of these contentions have at least a kernel of truth to them.
One overarching complaint is the choice of variables: which ones to include and which to omit? Why those variables and not others? Indeed, the selection of variables may lead economists to infer causation, especially if the data is only incidentally correlated. This leaves many to suggest that researcher bias must be at play in attaining and interpreting modelling results.
American economist Robert Lucas Jr decries the oversimplification of econometric models, especially as they are most often used to influence - if not outright set monetary policy. His chief complaint is that, limited as these models are, they don't allow for any changes economic actors might make.
While these criticisms tend to be very specific, the Austrian School of Economics has rejected econometrics altogether because the gathered data reflects past behaviour, which has proven to be an inexact predictor of future economic behaviour.
The last two global economic crises - in 2008 and now, amidst the global pandemic, certainly prove that point, right?
Whether you lean more toward the critical eye or marvel at its utility. econometrics is now firmly entrenched in global financial policy. Indeed, it even steers policymaking, at least to an extent. And, while you may have no desire to engage in the world of high finance or economics, it's good to know how the information governments use to make fiscal decisions is distilled.
If only for that reason, it would pay for everyone to seek a broader introduction to econometrics.